Derelict SFML destructor crashes

[Jeremy DeHaan]

On Tuesday, 18 December 2012 at 01:24:43 UTC, Mike Parker wrote:

> No, that's a bad idea. What if the C library you are using 
> maintains some sort of internal bookkeeping related to the 
> pointer? What if it keeps a copy of the pointer internally? If 
> the GC collects your D-side pointer, how is the C library 
> supposed to know it's time to cleanup everything associated 
> with that pointer?

I definitely hadn't considered that. I was mainly talking about 
the CSFML library, but I think even then trying to add the 
pointers into the GC would cause the problems you mention. Thanks 
for all the insight!